Utilising Genetic Interaction Networks to Elucidate the Anticancer and Diabetogenic Activities

Discovered over 40 years ago, statins are one of the most prescribed drugs in the world that have saved millions of lives. Beyond their main cholesterol-lowering purpose, statins exert anticancer activity. However, statins have also shown diabetogenic action, a major concern because more than 200 million people take statins worldwide. The aim of this thesis is to elucidate the genetic mechanisms, specifically genetic, chemical genetic and conditional interactions, by which statins act on cancer and diabetes.

The target of statins is 3-hydroxy-3-methylglutaryl-CoA reductase (HMGCR), the rate-limiting enzyme in the well characterised mevalonate pathway integral to the synthesis of cholesterol that has several branches at farnesyl diphosphate (FPP) to other outcomes potentially affecting diabetes and cancer. My hypothesis is there are many genetic, chemical genetic and conditional interactions mediating the anticancer and pro-diabetogenic activities of statins. Since defining complex genetics may be achieved by building interactive gene networks utilising genome-wide deletion libraries that do not yet exist in human cells, I used the genetic model Baker’s yeast (Saccharomyces cerevisiae) in three genetic backgrounds (S288C, UWOPS87-2421, and Y55).

In Chapter 2, genetic and chemical genetic interactions with the mevalonate pathway were investigated via a genome-wide analysis of 25,800 double deletion strains treated with atorvastatin, each lacking a gene in the statin pathway (HMG1 or BTS1) and a second gene in the yeast genome. Atorvastatin-hypersensitive mutants were validated in serial dilution spot assays and examined in the context of a multi-layer network comprising genetic and physical interactions. Functional subnetworks (modules) in the multi-layer network were identified and evaluated for network centrality as well as pathway enrichment, which identified the importance of specific genes mediating actin, ageing, unfolded protein response (UPR) and autophagy. I propose a model whereby deregulated actin may inhibit endocytosis and induce UPR, resulting in autophagic cell death. I also identified combination therapies of statins with other compounds that may enhance the anticancer activity of atorvastatin.

In Chapter 3, genetic and chemical genetic interactions mediating the diabetogenic activity of atorvastatin were investigated via a genome-wide analysis in the background of the established yeast models of anorexia and obesity. I generated 51,600 triple deletion strains, each lacking either the TGL3 and TGL4 genes required for triacylglyceride (fat) degradation (the obese model) or the DGA1 and LRO1 triacylglyceride synthesis genes (the anorexia model) and a third gene in the yeast genome, and measured growth of these triple deletion strains in the presence and absence of atorvastatin. Atorvastatin-hypersensitive mutants were validated in serial dilution spot assays and examined in the 1 context of a multi-layer network comprising genetic and physical interactions. Functional subnetworks (modules) in the multi-layer network were identified and evaluated for network centrality as well as pathway enrichment, which confirmed the importance of specific genes involved in ER-to-Golgi vesicle transport, UPR and autophagy as buffering mechanisms in these lipotoxic yeast models. Furthermore, I propose that lipotoxicity itself is a mechanism for atorvastatin-induced insulin resistance. This may occur via accumulation of acetyl-CoA as well as fatty acids and other lipotoxic intermediates that induce insulin resistance. I also identified potential combination therapies of statin with other compounds that may reduce the diabetogenic activity of atorvastatin.

In Chapter 4, conditional genetic and conditional chemical genetic interactions mediating hypoxia-specific mechanisms were investigated to further understand the molecular basis by which atorvastatin could elicit anticancer activity in hypoxic tumours. I screened 12,900 single deletion and 12,900 double deletion strains with a statin-related query gene deletion (BTS1) in the presence and absence of hypoxia. Atorvastatin-hypersensitive single and double mutants were validated in serial dilution spot assays and examined in the context of a multi-layer network comprising genetic and physical interactions. Functional subnetworks (modules) in the multi-layer network were identified and evaluated for network centrality as well as pathway enrichment, which identified the importance of specific genes involved in mitophagy and ubiquitination for hypoxia-specific atorvastatin activity. I also identified potential compounds that may specifically enhance the anticancer activity of statins in hypoxic tumours.

In summary, my results comprise a novel integration of methods for characterising complex genetics using methods that include epistasis in genetic effects. My results reveal the genetic, chemical genetic and conditional regulation underlying the anticancer and diabetogenic activity of atorvastatin in yeast that identify novel combination therapies and molecular mechanisms to further investigate in human cells and clinical trials.